The ~10kHz dip which they said balances out when averaged, still has a large effect on the listening window and a slight effect on all other curves, including the predicted in-room response, so the level should have been brought up a bit even still. The score of the speaker is highly hindered by that dip.
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Kali Audio IN-8 Studio Monitor Review
- Thread starter amirm
- Start date
Indeed, could be an issue for many people. Personally I would put in a couple of PEQS to remedy that situation and check if it sounded better. Come to think of it, who was the magician with that awesome program that predicted the effect of PEQs?
@BYRTT I believe
Too many posts to like!! SO I am "liking" no more 
I can't find enough superlative for what happened here. Kudos to Kali Audio and Amir.
I do understand the logistics nightmare of testing speakers but I am dying to see how some "audiophile" minimonitors would fare.
Great Job people!
I can't find enough superlative for what happened here. Kudos to Kali Audio and Amir.
I do understand the logistics nightmare of testing speakers but I am dying to see how some "audiophile" minimonitors would fare.
Great Job people!
Had a dirty quick run after the serious massage to those voicecoils thanks again to amirm, think it looks a nice speaker as is and each their taste for graphs visual or acoustics, filter is more than a few PEQ and relative high Q so quistion is how sample to sample will pair up but it does clean nicely up on stuff in surface charts or polars.
Many people seem to have a fixation to on-axis dip that most waveguides/horns have. This is strange to me, because this has been a well known phenomenom in hifi for at least 30 years! Elliptic and square horns are a bit easier. KEF uses a special phase plug.
It is audible to humans perhaps only in extreme nearfield, less than 0.5m (2') But these speakers are not meant for desktop-nearfield monitoring!
A closer look in measurements shows that power response is practically a straight descending line and listening window is on par with most conventional domes. Nothing sort of a problem at "normal" listening distance.
Some measurements and info of horns and waveguides
http://lib.tkk.fi/Dipl/2010/urn100200.pdf
At high frequencies (from 5 to 12 kHz) the diffraction dominates the directivity. The conical waveguide used has a sharp edge at the mouth. The sharp edge can be seen as a impedance discontinuity, which causes diffraction. It is known from the theory that the diffraction can be seen as a new sound source [15]. In the far field, this new source is out of phase with the direct sound at certain frequency. This phenomenon is almost solely an on-axis frequency response problem, because the off-axis responses are fairly unaffected except the 15° response (Figure 11.3). Even then, the directivity problem arises if a flat on-axis response is desired. The wavelength in the diffraction problem frequency range is approximately 7 cm to 3 cm. The distance from the mouth of the waveguide to the tweeter dome is approximately 2,5 cm. The frequency that correspond a half wavelength of this length is 6800 Hz. As can be seen in the model directivity plot, the diffraction problem is not exactly at one frequency as the theory would suggest. Instead it is smeared to frequency range from 5 kHz to 12 kHz. There are two explanations for this. First explanation is related to the model geometry. The dome is not a point source. Therefore the distance from different parts of the dome to the mouth is not constant. Second explanation is that in theory a plane wave pressure field is assumed. In reality the pressure field is more complex (Figure 10.2) and plane wave approximation is not adequate. The shape of the pressure wave is also frequency dependent. Figure 10.2. Sound pressure around the waveguide at 10 kHz. The severity of the diffraction problem is emphasized by the axisymmetry (Figure 10.3) because the distance from the tweeter dome to the edge is equal in all azimuth angles. An asymmetrical waveguide would smear the diffraction problem to a broader frequency range.
http://www.zaphaudio.com/hornconversion.html
http://www.acousticfrontiers.com/2014113speaker-off-axis-response-forward-firing-cone-dome-speakers/
https://kimmosaunisto.net/WaveGuides/WaveGuides.html
It is audible to humans perhaps only in extreme nearfield, less than 0.5m (2') But these speakers are not meant for desktop-nearfield monitoring!
A closer look in measurements shows that power response is practically a straight descending line and listening window is on par with most conventional domes. Nothing sort of a problem at "normal" listening distance.
Some measurements and info of horns and waveguides
http://lib.tkk.fi/Dipl/2010/urn100200.pdf
At high frequencies (from 5 to 12 kHz) the diffraction dominates the directivity. The conical waveguide used has a sharp edge at the mouth. The sharp edge can be seen as a impedance discontinuity, which causes diffraction. It is known from the theory that the diffraction can be seen as a new sound source [15]. In the far field, this new source is out of phase with the direct sound at certain frequency. This phenomenon is almost solely an on-axis frequency response problem, because the off-axis responses are fairly unaffected except the 15° response (Figure 11.3). Even then, the directivity problem arises if a flat on-axis response is desired. The wavelength in the diffraction problem frequency range is approximately 7 cm to 3 cm. The distance from the mouth of the waveguide to the tweeter dome is approximately 2,5 cm. The frequency that correspond a half wavelength of this length is 6800 Hz. As can be seen in the model directivity plot, the diffraction problem is not exactly at one frequency as the theory would suggest. Instead it is smeared to frequency range from 5 kHz to 12 kHz. There are two explanations for this. First explanation is related to the model geometry. The dome is not a point source. Therefore the distance from different parts of the dome to the mouth is not constant. Second explanation is that in theory a plane wave pressure field is assumed. In reality the pressure field is more complex (Figure 10.2) and plane wave approximation is not adequate. The shape of the pressure wave is also frequency dependent. Figure 10.2. Sound pressure around the waveguide at 10 kHz. The severity of the diffraction problem is emphasized by the axisymmetry (Figure 10.3) because the distance from the tweeter dome to the edge is equal in all azimuth angles. An asymmetrical waveguide would smear the diffraction problem to a broader frequency range.
http://www.zaphaudio.com/hornconversion.html
http://www.acousticfrontiers.com/2014113speaker-off-axis-response-forward-firing-cone-dome-speakers/
https://kimmosaunisto.net/WaveGuides/WaveGuides.html
Last edited:
It is a trade-off, but its effect is still apparent in the in-room curve and especially in the listening window curve; so IMO they should have upped the level a bit; it for sure would have scored better in the preference rating.
It is practially impossible to eq that in a passive loudspeaker's xo. And besides, at 8-12kHz head/ear position will change response more than that. I still strongly disagree about the audibility of this!
I cant find feature to hang on/attach a txt file so here is visual in high resolution so you can capture use a free trace tool, to try the hard core one get curve converted to IR wav-file and load in a convolution engine or convolute a few tracks with that filter, for example Jriver can load the IR filter and if you set it to play to disk you have a filtered track for comparison a non filtered track, some trace tool set phase to zero so before filter is converted to IR ensure curve is added minimum phase (MP) wich is very important, VituixCAD is free and can do all of those task. Alternative smooth curve with for example psychoacoustic or VAR smoothing in REW as in lower example and find the few PEQ in reverse that linearize below curve either using REW EQ feature or member pos free Rephase come handy for that. Haven't these nice speakers myself so blank how it works but the clues are there to get them cold clinic precision anechoic
or use a few softer PEQ based below scheme, wish it was me have fun.
thewas
Master Contributor
- Joined
- Jan 15, 2020
- Messages
- 7,348
- Likes
- 18,904
As I had written in my old review the Kali (like most coaxial speakers as also my KEF LS50) are often better listened of-axis as radiation irregularities add up on axis due to the symmetry of the driver. Personally I would also have "filled up" per EQ that 8-10 k region a bit more which I actually do on my setup, but this is complaining on a very high level
Now, about the K3 distortion dominance, unlike for an amp measurement it is a sign of good loudspeaker chassis as it shows that its drive has been optimised to be symmetric, important is of course it is low enough to be not hearable.
Now, about the K3 distortion dominance, unlike for an amp measurement it is a sign of good loudspeaker chassis as it shows that its drive has been optimised to be symmetric, important is of course it is low enough to be not hearable.
I don't like when diffraction interferences are called resonances. In my vocabulary resonance is a time domain phenomenom (a frequency starts to oscillate), but instead an interference doesn't oscillate/ring.
https://en.wikipedia.org/wiki/Resonance
https://en.wikipedia.org/wiki/Wave_interference
But then I noticed this
https://journals.aps.org/prb/abstract/10.1103/PhysRevB.65.195406
In loudspeakers, a diffraction interference typically appears as narrowing directivity (dip in DI curve) and usually it makes a dip in on-axis response. All this because radiation pattern is a balloon and initial sound energy is spread in all dimensions instead of being more or less beamed towards the listener.
Here are my measurements of a SEAS T18 6½" coaxial in my living room. 1 meter dispersion with 4ms gating and 1,5 meter on-axis with 200ms gating which is enough to mask the on-axis dip.
https://en.wikipedia.org/wiki/Resonance
https://en.wikipedia.org/wiki/Wave_interference
But then I noticed this
https://journals.aps.org/prb/abstract/10.1103/PhysRevB.65.195406
In loudspeakers, a diffraction interference typically appears as narrowing directivity (dip in DI curve) and usually it makes a dip in on-axis response. All this because radiation pattern is a balloon and initial sound energy is spread in all dimensions instead of being more or less beamed towards the listener.
Here are my measurements of a SEAS T18 6½" coaxial in my living room. 1 meter dispersion with 4ms gating and 1,5 meter on-axis with 200ms gating which is enough to mask the on-axis dip.
Last edited:
Maybe Amir could be so kind to do also an MMM measurement in his garage (see Charles Sprinkle's video or MMM presentation here).
It would be interesting to compare the MMM response with LW measurement and PIR estimation (and no need to buy another software option
)
thewas
Master Contributor
- Joined
- Jan 15, 2020
- Messages
- 7,348
- Likes
- 18,904
I am always using MMM to room EQ my speakers, so here is the MMM response of both my Kali IN-8 in my not very big room vs. the ASR Klippel measurements:Maybe Amir could be so kind to do also an MMM measurement in his garage (see Charles Sprinkle's video or MMM presentation here).
It would be interesting to compare the MMM response with LW measurement and PIR estimation (and no need to buy another software option
As it can be seen the resemblance is not bad at all above the modal region (around 500 Hz).
Here also my quick'n'dirty sound power estimation measurement (also MMM but in a closer radius in all directions around the loudspeaker) vs. the Klippel measurement:
Imho not too shabby either considering these measurements take less than a minute, imho not a bad quick'n'dirty check of the overall responses.
- Joined
- Nov 22, 2018
- Messages
- 1,137
- Likes
- 1,510
First, take a close look at the horizontals in post 594, then try listening ~20 degrees off-axis before EQ*. I'd personally target any EQ efforts for that axis. EQ cannot fully resolve the on-axis symmetry issues, and starting with the flattest natural response is a plus. All that said, off-axis peaks are easily absorbed at such high frequencies, so different approaches could work.
* FWIW, Geddes optimized his round waveguides for 22 degrees. Coincidence?
Nice explanation. I just have to add that this has the interesting effect of making Genelec's Ones more impressive.
I now also understand that ME Geithain RL901K's design (https://www.me-geithain.de/en/rl-901k2.html) where the coaxiality is between the woofer and mids+tweeter combo combination is probably better than this, even if it doesn't look like so, instinctually, since all kinds of interference (including diffraction) due to the massive obstacle in front of the woofer shouldn't be a problem at those wavelengths (the woofer crosses at 550Hz -> ~60cm) and the midwoofer is so close to the tweeter that the vertical dispersion should be almost as good as a triple coaxial design.
thewas
Master Contributor
- Joined
- Jan 15, 2020
- Messages
- 7,348
- Likes
- 18,904
In theory yes, in reality the Geithain loudspeakers aren't more linear on- or off-axis, here the driver responses of a three way coaxial model without crossover
which of course need heavy EQing
to even reach this not so great linearity on-axis
and off-axis
Source: https://www.soundandrecording.de/eq...-nahfeldmonitor-mit-koaxialem-aufbau-im-test/
This doesn't mean that its bad, in contrary, I really like the voicing of the Geithain speakers, just today I listened and enjoyed their new 921K, same as I love the 901K, our ears are luckily not so sensitive to small peaks and dips.
Last edited:
While this is not wrong, the 944K isn't their flagship, and since the SAR measurements match really well those given for the 944K:
I guess we can trust those given for the 901K:
which are definitely cleaner. They don't seem to have the same design as well, where one has a plate in front of a low crossing woofer and the other has the midrange and tweeter stuck to their small siblings' grill.
Personally, my assumption with any coaxial is that its safer to listen a little bit off axis. Usually just a fewegrees is enough to clean up the frequency response issues, if they're audible at all (though they're a bit more audible in a neafieldish setup). Even KEFs sound notably better off axis, and we know Andrew Jones tends to design speakers this way. too
Coaxials aside, I'd estimate about half of the speakers I've measured have a smoother or slightly more linear response slightly off axis.
This is one thing not discussed in the Olive paper, which as far as I can tell, made no effort to find an ideal axis. This is understandable given the effort involved and the fact this wasn't necessary to develop a preference score, but it does probably bias the speakers that perform better on-axis somewhat.
I am surprised Kali doesn't say anything about positioning at all in their manual though.
The rather drastic difference in preference score with the listening window vs on-axis in this case (6.7 vs 7.4) makes me think we should maybe include a separate calculation with the listening window in the preference chart for each speaker. I'll bring this up in the preference score thread though to not get too off topic.
Similar threads
